Cryptosporidiumspecies are the most common cause of gastrointestinal illness in treated recreational water venues. In order to protect public health during swimming, Cryptosporidium-sized microsphere removals by high-rate sand filtration with six coagulants were evaluated with a 5.5 m3 pilot-scale swimming pool. A sand filter without coagulation removed 20–63% of Cryptosporidium-sized microspheres. Cryptosporidium-sized microsphere removals exceeded 98% by sand filtration with five of the six tested coagulants. Continuously feeding coagulants A, B, and F (i.e., organic polymers) led to coagulant accumulation in the system and decreased removals over time (<2 days). Coagulant E (polyaluminum chloride) consistently removed more than 90% of microspheres at 30 m/h while the removals dropped to approximately 50% at a filtration rate of 37 m/h. Coagulant C was a chitosan-based product that removed fewer microspheres compared with other products, <75%, under the studied conditions. Results indicated aluminum-based coagulants (coagulants D and E) had an overall performance advantage over the organic polymer based coagulants primarily in terms of their tendency not to accumulate in the water and cease to be effective at improving filter efficiency.
INTRODUCTION
Cryptosporidium species are the most common cause of gastrointestinal illness associated with recreational water venues (Juranek 1995; Marion et al. 2010) and numerous waterborne outbreaks of cryptosporidiosis have been linked to swimming pools in the United States, United Kingdom, Australia, etc. (LeChevallier et al. 1991; Lisle & Rose 1995; Public Health Laboratory Service (PHLS) 2000; Puech et al. 2001; Briancesco & Bonadonna 2005; Karanis et al. 2006; Lu et al. 2013).
For filtered and disinfected aquatic venues, Cryptosporidium is a public health threat, because it is extremely resistant to halogen disinfection and its small size makes it a challenge for filtration systems (Korich et al. 1990; Carpenter et al. 1999). The drinking water industry has demonstrated that Cryptosporidium removal throughout all stages of the classical treatment process is largely influenced by the effectiveness of coagulation pretreatment (Dugan et al. 2001; Amburgey 2002; Amburgey et al. 2004; Hankins et al. 2006; Cummins et al. 2010; Karim et al. 2010; Lopez et al. 2010). Particles tend to repel each other, and there is no natural tendency for them to attach to other particles or surfaces unless effective coagulation to destabilize the particles is performed (Edwards 1997; Gao et al. 2002). Inorganic trivalent metal ions (such as aluminum and ferric iron) and water-soluble organic polymer coagulants are widely used for particle and natural organic matter (NOM) coagulation (Polasek & Mutl 2002; Bolto & Gregory 2007). Charge neutralization and sweep coagulation are the predominant mechanisms for Cryptosporidium removal via coagulation (Butkus et al. 2003; Xagoraraki & Harrington 2004; Okuda et al. 2006). The largest removals in drinking water studies have been shown to occur at high alum doses where aluminum hydroxide precipitation was observed to be extensive (Xagoraraki & Harrington 2004). Since polymer addition generally does not impact the pH of the water being treated, pH adjustment is not necessarily required for optimal coagulation (Emelko & Huck 2003). Polymers acting as coagulants usually contain materials with high charge density (Bolto & Gregory 2007). Polydiallydimethyl ammonium chloride (polyDADMAC) is a commonly used water-soluble polymer. PolyDADMAC was found to be very effective for removing disinfection by-product (DBP) precursors and NOM when used as a primary coagulant or coagulation aid (Polasek & Mutl 2002; Chang et al. 2005; Hankins et al. 2006; Wei et al. 2010). Natural cationic polymers, such as chitosan, were reported to coagulate with particles and enhance the particle removals for drinking water applications (Guibal et al. 2006; Bolto & Gregory 2007; Fabris et al. 2010). Chitosan has a low charge density (Parsons et al. 2007). Cryptosporidium parvum oocyst removals by filtration following chitosan coagulation at optimal dosages were comparable to those achieved when filtration (at 10 m/h) was preceded by alum or iron coagulation during optimized operation in a drinking water pilot plant (Brown & Emelko 2009). The effectiveness of coagulation depends on dosage (Divakaran & Pillai 2001). Investigation indicated chitosan coagulation at a dosage less than 1.0 mg/L did not result in appreciable improvements in Cryptosporidium parvum oocyst removal (Brown & Emelko 2009).
Direct filtration is similar to conventional treatment in that a coagulant is used to form larger particles, but coagulated water is applied directly to the filters without a gravity separation step. Direct filtration is generally used for low and consistent turbidity water. The removal of Cryptosporidium in direct filtration is usually lower than that in conventional drinking water treatment (Nieminski et al. 1995; Nieminski & Ongerth 1995; Amburgey 2002). A two-year evaluation of Cryptosporidium was conducted at a full-scale treatment plant and a pilot plant operating with coagulation and direct filtration (Nieminski & Ongerth 1995). Results showed 99.87% (2.9 log) of Cryptosporidium was removed when the treatment plant produced water of consistently low turbidity (i.e., 0.1–0.2 NTU). There is minimal published research on Cryptosporidium removals from swimming pools by filtration with coagulation.
The primary objective of this study was to determine the approximate level of Cryptosporidium oocyst-sized microsphere removals that can be achieved in a swimming pool system through high-rate sand filtration with six coagulants. Removals of microspheres, particles, turbidity, and organic matter (indicated by UV254) were evaluated for each coagulant in addition to pressure loss across the filter.
MATERIALS AND METHODS
Experimental setup
Synthetic pool water
The recipe for the synthetic pool water was based on a previous swimming pool water quality investigation (Lu 2012). Chemical characteristics of the simulated swimming pool water samples were pH of 7.5, hardness of 200 mg/L, as CaCO3, alkalinity of 100 mg/L as CaCO3, and free chlorine of 2 mg/L. Approximately 5,500 L of Charlotte, NC tap water with dissolved organic carbon (DOC) of approximately 1 mg/L was supplemented with NaHSO4, CaCl2, NaHCO3, and Ca(OCl)2 to create the synthetic pool water. A synthetic bather load based on the major constituents of sweat and urine was shown to have no impact on the surface charge or particle removal and was not used in these experiments (Lu 2012).
Cryptosporidium-sized polystyrene microspheres and coagulants
The use of polystyrene microspheres as an oocyst surrogate has been done by multiple researchers and was also done in this study (Li et al. 1997; Amburgey 2002, 2011; Dai & Hozalski 2003; Amburgey et al. 2004, 2005, 2012). Microspheres with a diameter of 4.5 μm (Fluoresbrite™ Carboxylate YG microspheres, Cat. #16592, Polysciences, Inc., Warrington, Pennsylvania, USA) were used as the surrogate since these microspheres are virtually identical to Cryptosporidium oocysts in size, shape, density, and surface charge in pool water (Amburgey 2010). The concentration of the stock suspensions of microspheres was 4.37 × 1011 #/L. A diluted suspension of microspheres with a concentration of 4.37 × 108 #/L was prepared by 1:1,000 dilutions of the stock solution. Microsphere samples were mixed by vortexing and hand shaking for at least two minutes each before analyzing. Samples were passed through 3.0 μm pore size polycarbonate track-etched filters. Each polycarbonate filter was mounted on a glass microscope slide with a polyvinyl alcohol-DABCO solution, covered with a glass cover slip and counted under an epi-fluorescent microscope at 100× total magnification (Freer 1984; Harlow & Lane 1988; Arrowood 2002). For ease of counting and obtaining statistically valid data, microscope slides needed to contain between 10 and 150 microspheres. Removal efficiency was calculated by comparing the concentrations between influent and effluent samples. Six coagulants used in pool water treatment were selected. Coagulants A and B were polyDADMAC products, coagulant C was chitosan, coagulant D was an aluminum-based coagulant, and coagulant E was polyaluminum chloride (Robarb 2000; BioGuard 2001; SeaKlear 2008; Vantage 2009; Kemira 2012). The formulation of coagulant F was proprietary.
General experimental approach
Experiments were performed over multiple theoretical hydraulic detention times (or turnovers) of the swimming pool, which were 8 hours each. Coagulants were fed at one vendor-recommended dose every turnover. Extended feeding of coagulants was evaluated. Samples were taken during each turnover. High-rate sand filtration control experiments without coagulant addition were also conducted. A backwash was performed after each experiment along with replacing the water in the pool and the sand in the filter.
Absorbance versus coagulant A concentration calibration curve (10 mm cell).
Data analysis
Box plots were used to represent the results for each coagulant. A box plot is a convenient way of graphically depicting groups of numerical data through their five-number summaries: the smallest observation (sample minimum), lower quartile, median, upper quartile, and largest observation (sample maximum).
Quality assurance and quality control
A control experiment was conducted without filter media to test if there were microsphere losses in the system. An average of 1% removal (approximately zero) was obtained and demonstrated no significant system losses. Duplicate experiments were conducted and triplicate samples were taken. The swimming pool was drained, rinsed, refilled, and recirculated with tap water at least three times between experiments to limit any cross-contamination between experiments. Fresh sand was used for each experiment and was backwashed (at 55 m/h and 29 °C to obtain a bed expansion of approximately 30–40%) with simulated pool water for 5 minutes before experiments to ensure the sand was clean and the sand grains were restratified.
RESULTS AND ANALYSIS
Zeta titration for six coagulants into simulated swimming pool water (106 #/mL microspheres) (for coagulant D: 1 mg/L = 160.5 g/m2; for coagulant E: 1 mg/L as product = 0.1 mg/L as Al).
Zeta titration for six coagulants into simulated swimming pool water (106 #/mL microspheres) (for coagulant D: 1 mg/L = 160.5 g/m2; for coagulant E: 1 mg/L as product = 0.1 mg/L as Al).
A suspension of microspheres was used as a model colloid and coagulated by these cationic coagulants. The most likely mechanism of coagulation by these cationic coagulants is charge neutralization (Singley 1970; Bratby 2008; Letterman & Yiacoumi 2010).
Cryptosporidium-sized microsphere removals
Microspheres removal by sand filter during 24 hrs (1.8 #/mL microspheres, 30 cm sand, 37 m/h filtration rate).
Microspheres removal by sand filter during 24 hrs (1.8 #/mL microspheres, 30 cm sand, 37 m/h filtration rate).
Measured and estimated coagulant a concentration in each turnover
Time (hr) . | Influent concentration (mg/L)a . | Effluent concentration (mg/L)b . | Average influent and effluent (mg/L)c . | Estimated concentration (mg/L)d . | Differences between ‘c’ and ‘d’e . |
---|---|---|---|---|---|
8 | 1.40 | 1.38 | 1.39 | 1.56 | 12% |
16 | 2.76 | 3.03 | 2.90 | 3.12 | 7% |
24 | 4.36 | 4.41 | 4.39 | 4.68 | 6% |
32 | 5.86 | 6.07 | 5.97 | 6.24 | 4% |
40 | 7.36 | 7.72 | 7.54 | 7.8 | 3% |
48 | 9.16 | 9.09 | 9.12 | 9.36 | 3% |
Time (hr) . | Influent concentration (mg/L)a . | Effluent concentration (mg/L)b . | Average influent and effluent (mg/L)c . | Estimated concentration (mg/L)d . | Differences between ‘c’ and ‘d’e . |
---|---|---|---|---|---|
8 | 1.40 | 1.38 | 1.39 | 1.56 | 12% |
16 | 2.76 | 3.03 | 2.90 | 3.12 | 7% |
24 | 4.36 | 4.41 | 4.39 | 4.68 | 6% |
32 | 5.86 | 6.07 | 5.97 | 6.24 | 4% |
40 | 7.36 | 7.72 | 7.54 | 7.8 | 3% |
48 | 9.16 | 9.09 | 9.12 | 9.36 | 3% |
c: average of column a and b, which meant c=(a+b)/2.
e: difference between c and d, which meant e=(c-d)/[(c+d)/2]*100%.
Microsphere removals by high-rate sand filtration with coagulants, A, B, C, D* (for coagulant D: 1 mg/L = 160.5 g/m2), E, and F (duplicate experiments with triplicate samples).
Microsphere removals by high-rate sand filtration with coagulants, A, B, C, D* (for coagulant D: 1 mg/L = 160.5 g/m2), E, and F (duplicate experiments with triplicate samples).
Particle counts, turbidity, UV254, and pressure variation
(a) Particle counts, (b) turbidity, (c) UV254 transmittance, and (d) pressure variations over treatment time (1.56 mg/L coagulant A, 1.8 #/mL microspheres, 30 cm sand, 37 m/h filtration rate).
(a) Particle counts, (b) turbidity, (c) UV254 transmittance, and (d) pressure variations over treatment time (1.56 mg/L coagulant A, 1.8 #/mL microspheres, 30 cm sand, 37 m/h filtration rate).
Turbidity was another parameter used to evaluate the particle removal in this system. Turbidity variation for coagulant A is illustrated in Figure 6(b). Initial simulated pool water turbidity was approximately 0.2 NTU. Effluent turbidity for coagulant A was less than 0.2 NTU.
The approximate removal of organic matter was monitored by continuously measuring the UV transmittance at 254 nm. Influent and effluent UV254 transmittance are shown in Figure 6(c). Both filter influent and effluent water UV254 transmittance was in the range of 96% to 98% and showed no obvious trends. Similar UV254 variations were observed for the rest of the five coagulants (data not shown). There was no significant removal of UV254 in the filter system for these experiments. A previous study showed polymeric coagulants were not effective in reducing UV254 absorption (Freese et al. 2001).
Pressure losses over treatment time for six coagulants (1 recommended dosage of coagulant, 1.8 #/mL microspheres, 30 cm sand, 37 m/h filtration rate).
Pressure losses over treatment time for six coagulants (1 recommended dosage of coagulant, 1.8 #/mL microspheres, 30 cm sand, 37 m/h filtration rate).
Coagulation evaluation
Streaming current variation over treatment time (1.56 mg/L coagulant A, 1.56 mg/L coagulant B, 1.56 mg/L coagulant C, 305 g/m2 coagulant D, 0.1 mg·Al/L coagulant E, 1.56 mg/L coagulant F, 1.8 #/mL microspheres, 30 cm sand, 37 m/h filtration rate).
Streaming current variation over treatment time (1.56 mg/L coagulant A, 1.56 mg/L coagulant B, 1.56 mg/L coagulant C, 305 g/m2 coagulant D, 0.1 mg·Al/L coagulant E, 1.56 mg/L coagulant F, 1.8 #/mL microspheres, 30 cm sand, 37 m/h filtration rate).
DISCUSSION
PolyDADMAC based coagulant
Greater than 99% removal was achieved by polyDADMAC products, which indicated that polyDADMAC was a promising coagulant in Cryptosporidium removal from swimming pools. Similar performance for polyDADMAC coagulation of drinking water was reported previously in drinking water treatments (Bolto & Gregory 2007; Wei et al. 2009, 2010). However, the polyDADMAC product accumulated in the recirculating pool system after multiple turnovers, which would not occur in a drinking water system, due to the unique circumstance of swimming pool water treatment, where continuous circulation of pool water can lead to charge reversal. Charge reversal was also observed in the zeta potential titration results. The reversal of the zeta potential of the microspheres also indicated that the coagulant adsorbed onto the surface of the microspheres. A similar conclusion was made in the research on the interaction between Cryptosporidium and coagulants (Bustamante et al. 2001).
Chitosan
Chitosan is a polymer with a low charge density (Parsons et al. 2007). In acidic solutions, chitosan becomes an extended chain and charged, while in neutral solutions, chitosan is a more coiled structure and only slightly charged (Pan et al. 1999; Huang et al. 2000). When pH values shift from 4 to 7, the positive charge on the chitosan surface significantly decreases, and the contribution toward charge neutralization by chitosan to destabilize particles is less significant in neutral pH conditions (Huang & Yin 1996; Parsons et al. 2007). However, swimming pool water pH must be maintained around 7.5 for human comfort and equipment safety concerns. The optimum dosage for chitosan coagulation would be expected to be smaller in acidic solutions (pH <7) due to the increased number of protonated amine groups on chitosan at lower pH, and the destabilization of particles has been enhanced by the increased number of charged groups (Pan et al. 1999; Huang et al. 2000). In addition, a non-charge coagulation mechanism has been proposed for chitosan (Parsons et al. 2007), but the non-charge mechanism did not appear to result in significantly increased microsphere removals by chitosan coagulation under the conditions examined in the present study.
Coagulant C (chitosan) could only remove up to 80% of Cryptosporidium-sized microspheres. This can be related back to surface zeta potential, in that chitosan could not achieve the minimum target zeta potential, −10 mV, at the recommended dosage of 1.56 mg/L as product under the experimental condition. The microsphere removals by coagulant C (chitosan) were contradictory to the research conducted on in-line filtration (in a non-swimming pool water treatment system) with chitosan at pH 7.3 to 7.4, which showed approximately 99% Cryptosporidium and microsphere removal with 1.5 mg/L chitosan (Brown & Emelko 2009). Possible reasons for the differences in removal for chitosan between the two studies include: chitosan source, physical and chemical processing of the chitosan, chitosan concentration, coagulant demand, filtration rates, flocculation, and media size and depth. In particular, the filtration rate was 10.4 m/h compared with 37 m/h in this study.
Aluminum based coagulants
Alum has previously performed better than the polyDADMAC and chitosan on NOM removal from drinking water (Bolto 2001). However, filter pressure loss increased significantly compared to the other coagulants used in this study. With aluminum-based coagulants in the present study (i.e., coagulants D and E), there did not appear to be a tendency toward overdosing/charge reversal over time under the conditions studied, which was not the case for cationic polymer coagulants. However, removals by aluminum-based coagulants were sensitive to both the filtration rate and the degree of clogging of the filter. Previous investigations have also shown that sand filter removal efficiencies with aluminum based coagulants are also dose dependent (Croll et al. 2007; Goodman 2011).
CONCLUSIONS
Cryptosporidium-sized microsphere removals from a pilot-scale swimming pool via high-rate sand filtration with six coagulants were studied separately. Cryptosporidium-sized microsphere removals were 20–63% by filtration with high-rate sand filtration without coagulation. Up to 99% of Cryptosporidium-sized microspheres were removed through high-rate sand filtration with coagulants A, B, D, and F at 37 m/h. Continuously feeding a cationic organic coagulant (coagulants A, B, and F) led to coagulant accumulation and a positive surface charge of particles in the system, and consequent decreases in the average filter removals to less than 60% within 48 hours. Coagulant D was an aluminum-based coagulant and did not appear to cause charge reversal, but the removals declined over time as the filter clogged and the pressured drop across the filter increased. Stable microsphere removals of greater than 90% were achieved with coagulant E (polyaluminum chloride) only after decreasing the filtration rate to 30 m/h.
Coagulant C was a chitosan-based product that removed fewer microspheres compared with the other products (less than 75% under the studied conditions). Overall, aluminum-based coagulants (coagulants D and E) had a performance advantage over the organic polymer based coagulants primarily in terms of their tendency not to accumulate in the water after multiple rounds of coagulation/filtration and cease to be effective at improving filtration efficiency.
ACKNOWLEDGEMENTS
The authors wish to thank the support of the National Swimming Pool Foundation (NSPF), the National Natural Science Foundation of China (41403090), and Arch Chemicals, Inc. (now Lonza Group Ltd) and other companies whose products were tested.